Researchers at Singapore’s Nanyang Technological University (NTU Singapore) have been trying to use wearable devices to predict whether or not someone is likely to experience depression. To that end, the researchers have published a study in the JMIR mHealth and uHealth journal that suggests that there are indeed key physiological indicators that correlate with depression.
In the study, the researchers asked 290 adult participants to wear Fitbit Charge 2 devices for 14 straight days, taking them off only to shower and to recharge the battery. The Fitbits tracked physical activity, heart rate, energy levels, and sleep patterns, and then analyzed that data with a Ycogni machine learning model that was built to predict depression.
The results indicate that Ycogni was able to identify people with a high risk of depression and people with no risk of depression with 80 percent accuracy. Ycogni’s predictions were compared to two health surveys that participants filled out at the beginning and end of the 14-day period.
In terms of physiological predictors, the NTU Singapore researchers found that poor sleep was the most telling risk factor for depression. People whose heart rates fluctuated in the middle of the night were more likely to have symptoms of depression, as were those who went to bed at different times and generally had less regular sleeping patterns.
The researchers believe that their system could help serve as an early warning system for high-risk people moving forward. They noted that nearly one billion people now wear activity trackers on a regular basis, and they are planning to study factors like skin temperature and smartphone usage to further refine the Ycogni algorithm.
The US Army has tried to use Fitbits to identify symptoms of COVID-19, while Fitbit itself has released a Fitbit Sense watch that can analyze sweat to gauge the wearer’s stress levels. Scientists are also trying to use wearables to monitor and treat conditions like diabetes and Parkinson’s disease.
(Originally posted on FindBiometrics)